How Marketers Make Better Decisions With Campaign Measurement

I trust older data more than the newest. For a flight simulation software client, purchases are credited across channels on a 14-day click window, so the latest days are always still filling in. We only raise the return target if total revenue held within 10% while ROAS climbed, and never by more than about 15% at a time. Most bad calls come from reading last week as finished.

The default in marketing dashboards is to use last click attribution, where if your customer views your ad on social, visits your blog post and then finds your site through search and places an order, 100% of the credit will be attributed to that final search. While last click attribution is easy to measure, it’s regularly the wrong measure. Channels which create demand and initial interest from a customer get none of the conversion credit.

We’ve made this attribution mistake in the past, using a last click attribution model that indicated we should increase our spend on search (taking credit for all conversions) and reduce spend on content that created demand. Luckily we realised this and changed our approach before we’d made a substantial loss but it still cost us significantly over two quarters.

We changed our approach to weighting conversions, taking into account the position of each touchpoint in the journey and the length of time between touchpoint and final action. For example, the first organic search from the unaware customer receives 40% of the credit, for opening up the path to purchase; the blog post (viewed 9 days before the conversion) receives 25% credit for sustaining interest in the offering; while the retargeting ad, seen three times in the last week receives only 10% credit, nudging them to make their final decision, but a decision which had been influenced by everything before that.

When we stopped counting downloads and started tracking meetings booked, everything changed. Our LinkedIn roundtables booked three times more meetings with actual buyers than webinars did. That told us where to put our budget. Track what gets you in a room with decision-makers, not what gets people to click.

A dashboard I reviewed credited paid search with most of a client’s signups, and the budget was already leaning that way. We added one open-text question to the signup form, “How did you first hear about us?”, and read the answers for a month. A surprising number named a podcast and a partner’s newsletter, both of which the analytics had filed under “direct.” We kept the search budget and moved a modest amount into sponsoring those two sources, and signups from them rose in the following quarter. Customers can tell you where the journey began, which a click path often cannot, so I now treat self-reported answers as a required companion to any attribution report.

The channel that looks busy isn’t always the one doing the work.

PI firm in Virginia. Running LSA and branded paid search side by side.

Lead counts looked fine on both. We’d been treating them equally for eight months.

Then we pulled cost per signed case by channel instead of cost per lead.

LSA: $840 per signed case.

Branded paid: $2,380 per signed case.

Same budget. Same general category. Three-to-one difference in actual cost.

Moved $1,500 a month from branded paid into LSA.

The next quarter: same spend, four more signed cases.

Nobody had asked the question at the channel level. We’d been measuring the wrong thing the whole time.

When customers touch several channels, I prioritize qualified pipeline and revenue across the journey, while using channel metrics to diagnose where progress is being lost. I don’t add together revenue claimed by multiple platforms or treat the last click as proof of causation. In our published Appomate case study, Meta was producing leads, but sales feedback showed that most were poor fits. We linked the lead flow to the CRM and evaluated which leads became real sales conversations, then changed the qualifying questions and creative around those findings. That led to a more useful decision than pursuing the lowest possible cost per lead: optimize for buyer fit even if the headline lead count falls. My advice is to agree the business outcome and evaluation window before comparing channels, then use sales feedback to check whether the dashboard’s success matches what happens after the click.

Since I focus on persona 360 mapping and behaviour targeting, I have stopped asking “which channel won?” a long time ago. In B2B, a single purchase usually involves three to five decision makers, each with a different pain point, and they touch different channels at different stages for different reasons. Last-click attribution doesn’t just get the credit wrong. It hides who you’re actually persuading and why they decided to react. So the results I give the most attention to are the ones that show the right person moving forward down the funnel, not the channel that happened to be there at the end.

In practice, that means I map intent before I look at any performance data. For every account we define the personas, such as the finance director worried about cost and risk, the operational lead worried about time, and the founder worried about growth. Then we map each channel to the stage and the pain point it serves. Search and AI answers capture people who are already looking for a solution. LinkedIn and content create recognition with the people who will be in the room. Email and sales conversations convert. Each channel is judged on its own job: did it reach the right persona, did it move them a stage closer, and did that account later turn into pipeline? Integration matters more than any single channel, because buyers don’t experience your channels separately. They experience one story, told well or badly across all of them.

One example: on a B2B campaign, paid search looked like the clear winner on cost per lead, and the instinct was to move budget away from LinkedIn and content. When we judged leads by whether they came from target accounts and matched our personas instead of by volume, the picture flipped. Many of the “cheap” search leads were researchers and competitors, while the accounts that turned into opportunities had nearly always engaged with LinkedIn content beforehand, often weeks earlier and through someone other than the person who filled in the form. We kept the budget, tightened search targeting around buying-intent terms, and started measuring pipeline from target accounts. Lead volume went down and qualified pipeline went up.

The lesson I’d give any growth marketer: you can’t optimise what you haven’t defined. Know your lead, meaning who they are, what keeps them up at night and what they need to believe before they buy, and the right metrics become obvious. Without that, attribution is just accounting.

This is a huge challenge for agencies and marketeers at the moment. As the technology advances, the way we measure results is falling behind. Users are no longer reliant on Google. Discovery happens away from a website, with social media algorithms or AI search channels able to recommend products or services that the user doesn’t even know they’re in the market for yet. At Appius, we’ve seen traffic declines across almost every client site over the last 12 months. However, we’re seeing key events and key event rate improve across every primary traffic channel. Why? The discovery, research and decision making is happening away from the website. But when it’s time to convert, that user still navigates to your site to complete the journey.

Instead of reporting on traffic, impressions or clicks like we used to in the good old days, we’re reporting on overall website metrics. For some clients, we exclude paid channels to focus on owned media (organic search, direct, organic social, email, referral). However, in most examples we’re seeing the combined impact of all these channels improving as user journeys become more fragmented.

Of course, we report on AI mentions, citation, cited URLs, sentiment and the usual KPIs that help us understand how the other channels perform. But from experience, the metric the stakeholders care about defines what those marketing efforts delivered for them.

I’m Charles Liu, CEO and co-founder of Cubic Promote, where I oversee marketing and growth for our promotional products and uniforms business. We do not judge every marketing channel by the same result because some are designed to build visibility, while others are expected to convert.

For certain advertising platforms, the main goal is impressions. In our industry, clients often need to see the Cubic Promote brand more than once before they enquire, so those campaigns are about creating as many relevant touchpoints as possible.

For Meta and other performance-focused ads, we look much more closely at leads and enquiries because the goal there is conversion.

One change that helped us was stopping the habit of evaluating every campaign mainly by clicks. We started giving more weight to impressions on awareness campaigns because repeated exposure builds recognition, even if the customer converts later through another channel.

That has helped us make better budget decisions. Instead of cutting a campaign because it did not generate an immediate enquiry, we first ask what job that channel was meant to do in the customer journey.

When customers touch several channels before converting, I focus less on assigning perfect credit and more on identifying which channels consistently influence meaningful business outcomes. Your mileage will vary with how precise you can make that definition, but in my case I tend to compare platform metrics with qualified leads, conversion quality, branded demand, and assisted behavior across the journey. In one campaign, paid social looked weak under last-click reporting, so cutting it seemed obvious. Upon review of assisted conversions and subsequent branded search, we actually saw that people exposed to the campaign were arriving later through other channels and converting at a higher rate. We kept the channel but changed how we evaluated it, so I’d say that the decision improved overall performance because we stopped judging every tactic as if it had to close the sale itself.

Form fills are the easiest number to watch and often the least useful one. I put more attention on how many leads turn into qualified sales conversations within 90 days. A channel producing cheap leads that never reply to sales outreach is costing you money, even when its cost-per-lead looks impressive on a monthly report.

I usually look at Pipeline Quality by channel: qualified opportunities, average deal size, and how long each deal takes to close. A paid search campaign might deliver fewer leads than a broad social campaign, yet produce more closed revenue because those searchers already know what they need. One shift I recommend is reviewing channel performance only after sales has had time to work the leads. Judging a campaign in week two rewards volume, while judging it in month three rewards the buyers who actually matter to the business.

1. No one buys a first home off a single brochure. Most people view a listing or drive past a property on a Sunday and then discuss with friends and family before contacting an agent. So for me at CheapForexVPS the most influential touch points would be those where most influence is transferred i.e. not the click. For our traders this would be a comparison article on the web, a recommendation on a trading forum or a helpful response in a chat regarding latency. I would also look to see which channels are appearing in the customer journeys of customers who have stuck with us after 3 months. Most influence transferred = genuine influence = customers who are convinced to purchase (as opposed to those influenced by a discount on a product).

2. (old measure) was content ‘value’ to direct sign-ups. Thus, (old measure) was easy ‘cut’ for budget. Surprisingly, when I changed measure to Assisted Conversions and 90-day customer retention instead, picture changed dramatically. People who read our ‘setup’ guides before landing on a paid search ad for same product brought same people back to site and converted better. In short, content had more value than we gave it credit for. We kept to budget for content and even made cuts on very broad search terms that only brought in short term customers (i.e. 1 month). Importantly, our cost per retained customer decreased dramatically. This has one very valuable lesson for me. The channel that closes deal is not always the channel that earned deal.

As there are multiple channels for users to take within an auto/accident journey. It can be as simple as seeing a social ad, finding your website organically then coming back to your site weeks later to convert. We look at all of these touch points using a position based (u shape) attribution model along with our multi touch UX events. Allowing us to place weight on first touch of discovery and last touch of lead conversion. Once you can correlate these technical metrics to your marketing efforts you can start to understand how your products UX plays a factor with all of your traffic sources.

After changing our focus from top of funnel CTR to time to complete session we noticed our performance ads were bringing high amounts of traffic to our website. Our users were taking too long to navigate our product to complete a claim and dropping off. After improving our digital product we focused more on mid funnel ux to help our higher quality sources. This changed our conversion rate on web and lowered our acquisition cost on all forms of paid advertising.

Balancing the impact of multiple marketing channels requires understanding which touchpoints resonate most with your specific audience. From my experience as a professional in skin health and aesthetics, tracking client conversions often reveals unexpected trends. For example, I once shifted from weighing campaign performance solely on direct bookings to analyzing post-treatment consultations. This change highlighted that clients exposed to educational content about laser treatments whether through newsletters or Instagram reels—were more likely to invest in multi-treatment packages. Why? Because education built trust and reduced fear about advanced technologies.

This insight reshaped our strategy. Instead of pouring resources into visual promotions of single treatments, we began focusing on client education campaigns that mapped solutions to specific skin concerns. The result? A 23% increase in package bookings over six months. My advice to others would be to gather data on why clients choose you and leverage that knowledge to align campaigns to their priorities, whether it’s post-treatment care tips or ingredient science breakdowns. Thoughtful communication earns trust, and trust drives decisions — not just for today but across every channel touchpoint.

With B2B, almost nobody converts off a single touch. A buyer might see a few posts from our team on LinkedIn, read a blog article, sit in on a webinar, and only then book a demo. If you only credit the last thing they clicked, you end up pouring money into whatever happens to sit closest to the form, which is usually organic/paid search or a direct visit, and you starve the things that actually built the interest. So I look at which channels show up most often in the journeys of people who became real opportunities, in particular the responses to ‘how did you hear about us?’ on our forms. This helped us gain better visibility on those ‘Direct’ traffic prospects. For us, they were coming from our podcast and LLMs like ChatGPT, giving me clear priorities for increased investment.

We give the most weight to the channel that started the conversation, and we record both the first and the last touch for every lead. A real example: our analytics said about 70% of visitors were “direct,” so Instagram looked weak next to search. Before moving budget to ads, we checked why, and found that Instagram’s in-app browser hides where visitors come from. That changed our decision: instead of judging Instagram by those numbers, we started tagging the real origin of every visit, saving it with each booking, and using a tracked link in the bio. Now we compare channels by booked calls, not by traffic.

“When customers interact with several marketing channels before making a purchase, I think one of the biggest mistakes is giving all the credit to whichever channel got the final click.

Someone might discover a product through an influencer on Instagram, look it up a few days later, and eventually purchase through a Google search. If you’re only looking at the last interaction, you’re missing a big part of what actually influenced that decision.

I think the key is understanding what each channel is contributing. Some are great at introducing people to a brand, while others are better at getting them to take action. You need to look at the whole picture before deciding what’s working and where your budget should go.

One example from our work is a campaign we ran for Ricola. Beyond generating 26 million impressions, we integrated retail tracking through MikMak and recorded more than 62,500 clicks to retail. That gave us a much more useful way to evaluate the campaign than reach alone.

For me, that’s where better marketing decisions start. It’s not about which channel has the most impressive numbers on paper. It’s about understanding what those numbers mean for the business and using that information to make smarter investments.”

Every channel needs to be judged against its own decision window. A webinar program was about to be cut at the end of a quarter because it showed almost no direct conversions at the two-week mark. The sales team’s notes told a different story, so we widened the evaluation to ninety days and traced the opportunities that had a webinar attendee in the buying group. Many deals that closed late in the quarter had started with an attendee, so the program was kept and its report was reset to a ninety-day view. Matching the review period to how long customers actually take to decide is a simple adjustment that prevents a lot of premature cuts.

Some of the most undervalued results sit in content that gets reused across several channels, so I track a Source Asset’s full footprint. Our team took an eight-part PPC video series and turned it into three long-form articles, 12 social snippets, and a gated PDF guide within two business days. Done manually, that work would have taken an estimated 25 to 30 staff hours.

Scoring that video series on views alone would ignore everything built from it. A buyer might find one of the articles through search, see a snippet on LinkedIn a week later, then download the guide. Every one of those touchpoints traces back to the same original recording. A useful approach is to tag repurposed pieces with their source in your analytics, so you can roll performance up to the original asset. That shows which raw material deserves more investment, which a channel-by-channel report will never reveal.

I look at incremental lift, not last click. Last click tells you who closed the deal, not who actually brought the customer in the door.

At Gazelle we sell trade-ins, so someone might see a paid social ad, search our brand a week later, then convert through email. If I only credited the email, I’d starve the channel that started the whole thing.

We ran holdout tests on our upper funnel spend and found a channel our dashboards made look weak was actually driving a huge chunk of our organic search growth. Once we started measuring it by incremental revenue instead of direct conversions, we increased budget there and total conversions went up across the board.

Now every channel gets judged by what happens when we turn it off, not just what shows up in the last touch.

I give the most attention to content that can be connected to pipeline and closed business. For a crypto compliance SaaS client, we connected HubSpot with Google Analytics to trace leads, deals and revenue back to more than 100 content pieces published over a year. That changed the evaluation from which pages attracted visits to which assets contributed to a commercial outcome. One downloadable report generated more than 50 new deals, giving us a concrete example of value that a traffic-only ranking would not explain. The better decision was to use that attribution evidence when prioritising content, while retaining visibility of the earlier touches instead of assigning all credit to the final click.

When customers touch several channels, I stop crediting whichever one collected the sale and ask which channels shortened or enabled the path to purchase. Last-click reporting rewards the channels closest to checkout, so it can make early-stage work look useless.

We changed how we evaluated our YouTube tutorials. Under last-click, they drove only 3% of sales, and we nearly cut the budget. When we looked at assisted conversions instead, 29% of all purchases included a tutorial view earlier in the journey. Customers who watched one converted 2.3x more often, and their average order was 18% larger.

So we kept the tutorials and doubled the number published, funding it with 15% of our retargeting budget. Over two quarters, total revenue grew 12%, and cost per new customer fell from $47 to $39.

The lesson: judge each channel by the role it plays in the journey, not by who was standing at the finish line.

I’m Mariia Grachova, CEO of Loona Agency. Over the past two years I’ve helped clients generate $51.5M in revenue.

Most of our clients are US contractors, where a $28,000 roof can take weeks or months between the first search and the signed contract. The buyer touches several channels on several devices, and by the time they call, GA4 often says “Direct”. So the results I pay attention to are the ones I can trace to money in the CRM: which query and page brought the first touch, whether it turned into a call or a form, and what the deal was worth. Analytics counts actions, the CRM counts money, and I trust the line that connects them.

Rapid Roofing on Long Island is a good example. When we started, calls (60 to 80% of their inquiries) from five Google Business Profile locations landed in an unassigned channel, and leads from AI search showed up as referrals. In the usual reports impressions grew 422% and looked like a triumph, while CTR fell from 0.17% to 0.08% and looked like a failure. Neither number told us where the money came from. We saved the first touch in a first-party cookie, gave each GBP location its own tracking number and pulled the deal amount from the CRM. That moved our effort away from broad queries and toward city pages built around specific intent. Leads grew 61% annualized and ROMI reached 10x in high season.

In our business almost nobody buys after a single touchpoint. A training head considering a simulator from Tecknotrove might read one of our articles in an industry magazine, visit our website, meet us at an exhibition, and months later search our name and send an enquiry. If you only credit the last click, branded search wins every time and everything that actually built the interest looks worthless.

So I give the most attention to the channels that start conversations, not just the ones that close them. Our editorial articles in international mining, aviation and defence trade publications are a good example. Measured by referral clicks alone, they looked weak, since very few readers click straight from a magazine article to our site. When we started asking new enquiries where they first came across us, and watching branded search interest after each article ran, these articles turned out to be where many conversations really began. That changed our decision from questioning whether to continue the articles to planning them as a regular part of the calendar.

The habit I’d recommend is to ask every new enquiry how they first heard of you, then compare that with what your analytics says. When the two disagree, the customer is usually right.

When a customer touches four or five channels before they buy, I’ve stopped asking which channel deserves the credit and started asking which one we would miss if we switched it off. The second question is the one that should move budget, and most reporting is built to answer the first.

What changed my mind was reconciling ad platform reports against what clients’ own CRMs and order systems recorded. We did that across 240 accounts earlier this year. If you add up the conversions that Google, Meta and LinkedIn each claimed for the same period, the total came out 34% higher than the sales that took place, because every platform counts any sale its ad came near. Before that exercise we were moving money towards whichever dashboard looked best that month, which in practice meant rewarding the platform with the most generous attribution window.

So the decision rule now is backend numbers first and holdout tests on branded search and retargeting, since those are the two channels most likely to be claiming sales that would have happened anyway. Where a holdout shows little lift, the money goes into prospecting that reaches people who have never heard of the business, which last-click reporting has undervalued for years. I’ll admit the holdout weeks make clients nervous, and I understand why, because nobody enjoys switching off a campaign that looks profitable just to find out whether it was.